Apart from Resolv being hacked, this type of DeFi vulnerability has appeared four times before

Original title: DeFi Has Seen Resolv's $25M USR Exploit Many Times Before
Original author: Camila Russo
Original source: https://thedefiant.io/news/hacks/defi-has-seen-resolv-s-usd25m-usr-exploit-many-times-before
Reprinted: Mars Finance
On a quiet Sunday morning, someone turned $100,000 into $25 million in about 17 minutes.
The target is the yield-bearing stablecoin protocol Resolv. Before Resolv paused the contract, its USR stablecoin, pegged to the US dollar, had fallen to a few cents. As of the time of this writing, USR remains severely unpegged, trading at around $0.25, with a decline of over 70% this week.
The ripple effect extends far beyond Resolv itself. Fluid/Instadapp absorbed over $10 million in bad debts in a single day, while facing over $300 million in net outflows on the same day, marking its largest single-day outflow on record. 15 Morpho vaults were affected. Euler, Venus, Lista DAO, and Inverse Finance subsequently suspended USR-related markets.
The mechanism that led to the spread of losses from this vulnerability—pricing unpegged stablecoins at $1 in the lending market—is not new. This has occurred at least four times in the past 14 months.
How the vulnerability works
The minting of USR follows a two-step off-chain process: users deposit USDC through the `requestSwap` function, and a privileged off-chain signing key `SERVICE_ROLE` finalizes the amount of USR issued through `completeSwap`.
The contract has a minimum output limit but no maximum limit. The keyholder signs what the contract executes.
The attacker gained access to the key through Resolv's AWS Key Management Service. They made two USDC deposits totaling around $100,000 to $200,000 and then used the stolen key to authorize the minting of 80 million USR in return. On-chain data shows the two transactions were for 50 million USR and 30 million USR, both completed within minutes.
'The Resolv USR vulnerability is not a bug—it's a feature that operates normally by design. That's the problem,' said on-chain analyst Vadim (@zacodil).
SERVICE_ROLE is a standard external account address, not a multi-signature. The admin key is protected by multi-signature, but the minting key is not.
'Resolv underwent 18 audits,' Vadim said, 'and one of the findings was directly titled 'Missing Cap.'
The attacker methodically exited: first converting the minted USR to wstUSR (the staked wrapped version) to mitigate market impact, then swapping it for ETH through Curve, Uniswap, and KyberSwap. The attacker's wallet held around 11,400 ETH (approximately $24 million). The ETH and BTC collateral pools supporting the entire system remained intact while the stablecoins collapsed.
How infections spread
The Resolv vulnerability is actually a combination of two incidents. The first was a minting vulnerability, and the second was a failure of the chain lending market.
When USR and wstUSR collapse, every lending market that accepts them as collateral faces the same issue: their oracles still price wstUSR at close to $1.
Risk analysis firm Chaos Labs founder Omer Goldberg documented this mechanism. His core finding is: 'The oracles are hardcoded, so they have never been repriced. wstUSR was marked at $1.13, while its trading price on the secondary market was around $0.63.'
Traders purchased wstUSR at a low price on the open market and then used it as collateral on Morpho or Fluid at an oracle price of $1.13, borrowing USDC and exiting.
At Fluid, the team raised short-term loans to cover 100% of bad debts and promised to fully compensate every user. At Morpho, co-founder Paul Frambot stated that about 15 vaults had significant exposures, all within high-risk, long-tail collateral strategies.
Notable curator Gauntlet stated, 'Several high-yield vaults have limited exposure.'
But D2 Finance directly refuted this claim, releasing on-chain data showing that Gauntlet's flagship 'USDC Core vault' allocated $4.95 million to the wstUSR/USDC market. Goldberg subsequently stated that Gauntlet's vault accounted for 98% of the market's lending liquidity.
Frambot stated in a written response to The Defiant: 'We have been researching how to present various risks more comprehensively. However, we do not think the core issue here is a lack of labeling.'
Frambot added: 'Morpho is oracle-agnostic, which means it allows curators to choose any oracle they believe is most suitable for a specific market. Morpho is open, permissionless infrastructure designed to outsource risk management to curators.'
'It is difficult to enforce an objectively 'correct' guardrail in all scenarios,' Frambot said. 'Imposing constraints at the protocol level also carries the risk of hindering legitimate strategies.'
Although the underlying protocol leaves risk management to the curators, some industry insiders believe that curators have not fulfilled their responsibilities.
'I believe the design of the curator industry is flawed because there is no real curation happening,' Marc Zeller stated on X.
As of the time of publication, Resolv, Gauntlet, and Fluid had not responded to The Defiant's request for comment.
A recurring failure mode
This is not a new type of attack. In January 2025, Usual Protocol's USD0++ was hardcoded by the curator MEV Capital as $1 in the Morpho vault.
Usual subsequently and suddenly adjusted the redemption floor price to $0.87 without any warning, causing lenders to be locked in the MEV Capital vault, which saw its utilization soar to 100%.
In November 2025, Stream Finance's xUSD collapsed after curators routed USDC deposits into a leveraged loop backed by this synthetic stablecoin. When its oracle refused to update, an estimated $285 million to $700 million in assets on Morpho, Euler, and Silo were at risk.
Moonwell experienced two oracle failures consecutively in October and November 2025, resulting in over $5 million in bad debts.
What this means for the curator model
Morpho's architecture outsources all risk decisions to third-party 'curators', who build the vaults, select collateral, set loan-to-value ratios, and choose the oracles. This theory posits that professional institutions possess deeper expertise, and competition can lead to better risk management, with the protocol responsible for enforcing the rules.
But curators rely on the yield generated to earn fees, creating an incentive to accept higher-risk, higher-yield collateral (such as yield-bearing stablecoins). The problem is that when these stablecoins become unpegged, the losses are borne by depositors, not the curators.
In the Resolv incident, some curators' automated bots continued to inject funds into the affected vaults hours after the vulnerability occurred, exacerbating the losses.
The reason for employing hardcoded oracles for yield-bearing stablecoins is to prevent short-term fluctuations from triggering unnecessary liquidations. However, this protection is only effective when stablecoins remain stable.
On-chain analytics firm Chainalysis stated in a post-mortem review that real-time on-chain detection capabilities are needed.
'The on-chain smart contracts are operating normally. The problem clearly lies in the broader system design and off-chain infrastructure,' the analytics firm stated.